Compare_Augmentation_VS_NonAugmentation.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "v5wp7G-e4vH7"
      },
      "outputs": [],
      "source": [
        "#import necessaries\n",
        "import tensorflow as tf\n",
        "from tensorflow import keras\n",
        "from tensorflow.keras.datasets import mnist\n",
        "from tensorflow.keras.preprocessing.image import ImageDataGenerator\n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "from tensorflow.keras.models import Sequential\n",
        "from tensorflow.keras.layers import Dense, Flatten\n"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# MNIST dataset loaded for training and testing\n",
        "(x_train, y_train), (x_test, y_test) = mnist.load_data()\n"
      ],
      "metadata": {
        "id": "mhtCBoOH4z1q"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Normalize and Reshape the input data# Normalize the input data\n",
        "x_train = x_train.reshape((60000, 28, 28, 1)) / 255.0\n",
        "x_test = x_test.reshape((10000, 28, 28, 1)) / 255.0\n",
        "\n",
        "# Convert the labels to one-hot encoding.One-hot encoding is a process of converting categorical data into a binary vector where only one element is set to 1, and all other elements are set to 0.\n",
        "y_train = keras.utils.to_categorical(y_train, 10)\n",
        "y_test = keras.utils.to_categorical(y_test, 10)\n"
      ],
      "metadata": {
        "id": "6HBjc_ZO40_s"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Show one image for each of the digits 0 to 9\n",
        "fig = plt.figure(figsize=(15, 2))\n",
        "print(x_train.shape)\n",
        "for i, image in enumerate(x_train[:10]):\n",
        "    # Create a new subplot for each image\n",
        "    plt.subplot(1, 10, i+1)\n",
        "    plt.xticks([])\n",
        "    plt.yticks([])\n",
        "    plt.grid(False)\n",
        "    plt.imshow(image, cmap=plt.cm.binary)\n",
        "    plt.xlabel(y_train[i])\n",
        "\n",
        "\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 202
        },
        "id": "7SWTPOTd45Kj",
        "outputId": "34c13ff6-f102-4b6d-cec0-9302ae3688b5"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "(60000, 28, 28, 1)\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.9/dist-packages/matplotlib/text.py:1279: FutureWarning: elementwise comparison failed; returning scalar instead, but in the future will perform elementwise comparison\n",
            "  if s != self._text:\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1500x200 with 10 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Define the model architecture\n",
        "model = Sequential([\n",
        "    Flatten(input_shape=(28, 28, 1)),\n",
        "    Dense(128, activation='relu'),\n",
        "    Dense(64, activation='relu'),\n",
        "    Dense(32, activation='relu'),\n",
        "    Dense(10, activation='softmax')\n",
        "])\n"
      ],
      "metadata": {
        "id": "CEH--ugU4_lU"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Compile the model\n",
        "model.compile(optimizer='adam', loss='categorical_crossentropy', metrics=['accuracy'])\n"
      ],
      "metadata": {
        "id": "KMDONUAc5B2s"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Before Augmentation\n",
        "\n",
        "model.fit(x_train, y_train, epochs=10, batch_size=128)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "7JgtonKwfH2d",
        "outputId": "d04bd4e2-f712-46fb-8228-2232b7ba3459"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/10\n",
            "469/469 [==============================] - 4s 6ms/step - loss: 0.3449 - accuracy: 0.9006\n",
            "Epoch 2/10\n",
            "469/469 [==============================] - 3s 7ms/step - loss: 0.1283 - accuracy: 0.9625\n",
            "Epoch 3/10\n",
            "469/469 [==============================] - 2s 5ms/step - loss: 0.0896 - accuracy: 0.9735\n",
            "Epoch 4/10\n",
            "469/469 [==============================] - 3s 5ms/step - loss: 0.0685 - accuracy: 0.9793\n",
            "Epoch 5/10\n",
            "469/469 [==============================] - 2s 5ms/step - loss: 0.0530 - accuracy: 0.9838\n",
            "Epoch 6/10\n",
            "469/469 [==============================] - 3s 7ms/step - loss: 0.0416 - accuracy: 0.9873\n",
            "Epoch 7/10\n",
            "469/469 [==============================] - 3s 6ms/step - loss: 0.0351 - accuracy: 0.9893\n",
            "Epoch 8/10\n",
            "469/469 [==============================] - 3s 6ms/step - loss: 0.0301 - accuracy: 0.9904\n",
            "Epoch 9/10\n",
            "469/469 [==============================] - 3s 5ms/step - loss: 0.0255 - accuracy: 0.9916\n",
            "Epoch 10/10\n",
            "469/469 [==============================] - 2s 5ms/step - loss: 0.0218 - accuracy: 0.9928\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<keras.callbacks.History at 0x7fd5f41cbbb0>"
            ]
          },
          "metadata": {},
          "execution_count": 9
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Before Augmentation\n",
        "test_loss, test_acc = model.evaluate(x_test, y_test)\n",
        "print(x_train.shape)\n",
        "\n",
        "print('Test accuracy before augmentation:', test_acc)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "u8g6Eqi-fNnT",
        "outputId": "4744d539-22b0-490e-d658-146bbcda9773"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "313/313 [==============================] - 1s 2ms/step - loss: 0.0808 - accuracy: 0.9798\n",
            "(60000, 28, 28, 1)\n",
            "Test accuracy before augmentation: 0.9797999858856201\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Augmentation parameter generated\n",
        "datagen = ImageDataGenerator(\n",
        "    rotation_range=10,\n",
        "    width_shift_range=0.1,\n",
        "    height_shift_range=0.1,\n",
        "    shear_range=0.1,\n",
        "    zoom_range=0.1,\n",
        "    fill_mode='nearest'\n",
        ")"
      ],
      "metadata": {
        "id": "NqKsMdqffUUL"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Fit the train data with Augmentation\n",
        "datagen.fit(x_train)\n",
        "augmented_images = []\n",
        "for batch in datagen.flow(x_train, batch_size=len(x_train), shuffle=False):\n",
        "    augmented_images.append(batch)\n",
        "    if len(augmented_images) * len(batch) >= len(x_train) * 2:\n",
        "        break\n",
        "augmented_images = np.concatenate(augmented_images)\n",
        "y_train_augmented = np.concatenate((y_train, y_train), axis=0)\n",
        "\n",
        "\n"
      ],
      "metadata": {
        "id": "9GpKwHdw5ELr"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Fit the testing data with Augmentation\n",
        "datagen.fit(x_test)\n",
        "augmented_test_data = []\n",
        "for batch in datagen.flow(x_test, batch_size=len(x_test), shuffle=False):\n",
        "    augmented_test_data.append(batch)\n",
        "    if len(augmented_test_data) * len(batch) >= len(x_test) * 2:\n",
        "        break\n",
        "augmented_test_data = np.concatenate(augmented_test_data)\n",
        "y_test_augmented = np.concatenate((y_test, y_test), axis=0)"
      ],
      "metadata": {
        "id": "5TTb1RSnBxTN"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "b_test_acc = test_acc"
      ],
      "metadata": {
        "id": "SlRZBPG39q8L"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# After Augmentation\n",
        "\n",
        "model.fit(augmented_images, y_train_augmented, epochs=100, batch_size=128)\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "dTvam6ZE5Rrd",
        "outputId": "febcdd09-e4c7-44ee-80ca-d74b6b8b0144"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/100\n",
            "938/938 [==============================] - 4s 4ms/step - loss: 0.2585 - accuracy: 0.9230\n",
            "Epoch 2/100\n",
            "938/938 [==============================] - 4s 4ms/step - loss: 0.1370 - accuracy: 0.9576\n",
            "Epoch 3/100\n",
            "938/938 [==============================] - 5s 6ms/step - loss: 0.1063 - accuracy: 0.9669\n",
            "Epoch 4/100\n",
            "938/938 [==============================] - 5s 5ms/step - loss: 0.0884 - accuracy: 0.9722\n",
            "Epoch 5/100\n",
            "938/938 [==============================] - 4s 5ms/step - loss: 0.0746 - accuracy: 0.9761\n",
            "Epoch 6/100\n",
            "938/938 [==============================] - 6s 6ms/step - loss: 0.0645 - accuracy: 0.9793\n",
            "Epoch 7/100\n",
            "938/938 [==============================] - 4s 5ms/step - loss: 0.0570 - accuracy: 0.9815\n",
            "Epoch 8/100\n",
            "938/938 [==============================] - 6s 6ms/step - loss: 0.0492 - accuracy: 0.9838\n",
            "Epoch 9/100\n",
            "938/938 [==============================] - 5s 5ms/step - loss: 0.0441 - accuracy: 0.9855\n",
            "Epoch 10/100\n",
            "938/938 [==============================] - 5s 6ms/step - loss: 0.0393 - accuracy: 0.9868\n",
            "Epoch 11/100\n",
            "938/938 [==============================] - 5s 6ms/step - loss: 0.0346 - accuracy: 0.9882\n",
            "Epoch 12/100\n",
            "938/938 [==============================] - 5s 5ms/step - loss: 0.0339 - accuracy: 0.9884\n",
            "Epoch 13/100\n",
            "938/938 [==============================] - 6s 6ms/step - loss: 0.0273 - accuracy: 0.9906\n",
            "Epoch 14/100\n",
            "938/938 [==============================] - 5s 6ms/step - loss: 0.0284 - accuracy: 0.9899\n",
            "Epoch 15/100\n",
            "938/938 [==============================] - 6s 6ms/step - loss: 0.0252 - accuracy: 0.9914\n",
            "Epoch 16/100\n",
            "938/938 [==============================] - 5s 5ms/step - loss: 0.0240 - accuracy: 0.9917\n",
            "Epoch 17/100\n",
            "938/938 [==============================] - 5s 5ms/step - loss: 0.0224 - accuracy: 0.9923\n",
            "Epoch 18/100\n",
            "938/938 [==============================] - 6s 6ms/step - loss: 0.0221 - accuracy: 0.9925\n",
            "Epoch 19/100\n",
            "938/938 [==============================] - 5s 5ms/step - loss: 0.0196 - accuracy: 0.9935\n",
            "Epoch 20/100\n",
            "938/938 [==============================] - 6s 6ms/step - loss: 0.0207 - accuracy: 0.9929\n",
            "Epoch 21/100\n",
            "938/938 [==============================] - 5s 5ms/step - loss: 0.0177 - accuracy: 0.9940\n",
            "Epoch 22/100\n",
            "938/938 [==============================] - 6s 6ms/step - loss: 0.0175 - accuracy: 0.9941\n",
            "Epoch 23/100\n",
            "938/938 [==============================] - 5s 5ms/step - loss: 0.0158 - accuracy: 0.9946\n",
            "Epoch 24/100\n",
            "938/938 [==============================] - 4s 5ms/step - loss: 0.0171 - accuracy: 0.9941\n",
            "Epoch 25/100\n",
            "938/938 [==============================] - 6s 6ms/step - loss: 0.0146 - accuracy: 0.9951\n",
            "Epoch 26/100\n",
            "938/938 [==============================] - 4s 5ms/step - loss: 0.0184 - accuracy: 0.9939\n",
            "Epoch 27/100\n",
            "938/938 [==============================] - 4s 5ms/step - loss: 0.0147 - accuracy: 0.9951\n",
            "Epoch 28/100\n",
            "938/938 [==============================] - 6s 6ms/step - loss: 0.0126 - accuracy: 0.9958\n",
            "Epoch 29/100\n",
            "938/938 [==============================] - 4s 5ms/step - loss: 0.0161 - accuracy: 0.9947\n",
            "Epoch 30/100\n",
            "938/938 [==============================] - 6s 6ms/step - loss: 0.0142 - accuracy: 0.9951\n",
            "Epoch 31/100\n",
            "938/938 [==============================] - 5s 5ms/step - loss: 0.0148 - accuracy: 0.9950\n",
            "Epoch 32/100\n",
            "938/938 [==============================] - 5s 5ms/step - loss: 0.0145 - accuracy: 0.9951\n",
            "Epoch 33/100\n",
            "938/938 [==============================] - 6s 7ms/step - loss: 0.0123 - accuracy: 0.9956\n",
            "Epoch 34/100\n",
            "938/938 [==============================] - 5s 5ms/step - loss: 0.0126 - accuracy: 0.9958\n",
            "Epoch 35/100\n",
            "938/938 [==============================] - 6s 6ms/step - loss: 0.0144 - accuracy: 0.9952\n",
            "Epoch 36/100\n",
            "938/938 [==============================] - 5s 5ms/step - loss: 0.0125 - accuracy: 0.9959\n",
            "Epoch 37/100\n",
            "938/938 [==============================] - 5s 5ms/step - loss: 0.0098 - accuracy: 0.9967\n",
            "Epoch 38/100\n",
            "938/938 [==============================] - 6s 7ms/step - loss: 0.0152 - accuracy: 0.9951\n",
            "Epoch 39/100\n",
            "938/938 [==============================] - 4s 5ms/step - loss: 0.0100 - accuracy: 0.9966\n",
            "Epoch 40/100\n",
            "938/938 [==============================] - 6s 6ms/step - loss: 0.0132 - accuracy: 0.9955\n",
            "Epoch 41/100\n",
            "938/938 [==============================] - 5s 5ms/step - loss: 0.0103 - accuracy: 0.9966\n",
            "Epoch 42/100\n",
            "938/938 [==============================] - 5s 6ms/step - loss: 0.0122 - accuracy: 0.9959\n",
            "Epoch 43/100\n",
            "938/938 [==============================] - 5s 6ms/step - loss: 0.0096 - accuracy: 0.9968\n",
            "Epoch 44/100\n",
            "938/938 [==============================] - 5s 5ms/step - loss: 0.0111 - accuracy: 0.9963\n",
            "Epoch 45/100\n",
            "938/938 [==============================] - 6s 6ms/step - loss: 0.0120 - accuracy: 0.9962\n",
            "Epoch 46/100\n",
            "938/938 [==============================] - 5s 5ms/step - loss: 0.0100 - accuracy: 0.9967\n",
            "Epoch 47/100\n",
            "938/938 [==============================] - 6s 6ms/step - loss: 0.0107 - accuracy: 0.9967\n",
            "Epoch 48/100\n",
            "938/938 [==============================] - 5s 5ms/step - loss: 0.0107 - accuracy: 0.9966\n",
            "Epoch 49/100\n",
            "938/938 [==============================] - 5s 5ms/step - loss: 0.0108 - accuracy: 0.9963\n",
            "Epoch 50/100\n",
            "938/938 [==============================] - 5s 6ms/step - loss: 0.0099 - accuracy: 0.9968\n",
            "Epoch 51/100\n",
            "938/938 [==============================] - 4s 5ms/step - loss: 0.0088 - accuracy: 0.9973\n",
            "Epoch 52/100\n",
            "938/938 [==============================] - 5s 6ms/step - loss: 0.0105 - accuracy: 0.9965\n",
            "Epoch 53/100\n",
            "938/938 [==============================] - 5s 5ms/step - loss: 0.0098 - accuracy: 0.9968\n",
            "Epoch 54/100\n",
            "938/938 [==============================] - 4s 5ms/step - loss: 0.0078 - accuracy: 0.9975\n",
            "Epoch 55/100\n",
            "938/938 [==============================] - 6s 6ms/step - loss: 0.0094 - accuracy: 0.9968\n",
            "Epoch 56/100\n",
            "938/938 [==============================] - 4s 5ms/step - loss: 0.0116 - accuracy: 0.9963\n",
            "Epoch 57/100\n",
            "938/938 [==============================] - 5s 5ms/step - loss: 0.0093 - accuracy: 0.9970\n",
            "Epoch 58/100\n",
            "938/938 [==============================] - 6s 6ms/step - loss: 0.0086 - accuracy: 0.9972\n",
            "Epoch 59/100\n",
            "938/938 [==============================] - 5s 5ms/step - loss: 0.0082 - accuracy: 0.9974\n",
            "Epoch 60/100\n",
            "938/938 [==============================] - 6s 6ms/step - loss: 0.0097 - accuracy: 0.9969\n",
            "Epoch 61/100\n",
            "938/938 [==============================] - 5s 5ms/step - loss: 0.0079 - accuracy: 0.9974\n",
            "Epoch 62/100\n",
            "938/938 [==============================] - 6s 6ms/step - loss: 0.0091 - accuracy: 0.9971\n",
            "Epoch 63/100\n",
            "938/938 [==============================] - 5s 5ms/step - loss: 0.0072 - accuracy: 0.9977\n",
            "Epoch 64/100\n",
            "938/938 [==============================] - 5s 5ms/step - loss: 0.0096 - accuracy: 0.9967\n",
            "Epoch 65/100\n",
            "938/938 [==============================] - 6s 6ms/step - loss: 0.0081 - accuracy: 0.9975\n",
            "Epoch 66/100\n",
            "938/938 [==============================] - 5s 5ms/step - loss: 0.0099 - accuracy: 0.9969\n",
            "Epoch 67/100\n",
            "938/938 [==============================] - 6s 6ms/step - loss: 0.0071 - accuracy: 0.9977\n",
            "Epoch 68/100\n",
            "938/938 [==============================] - 5s 5ms/step - loss: 0.0090 - accuracy: 0.9973\n",
            "Epoch 69/100\n",
            "938/938 [==============================] - 5s 5ms/step - loss: 0.0092 - accuracy: 0.9972\n",
            "Epoch 70/100\n",
            "938/938 [==============================] - 6s 6ms/step - loss: 0.0071 - accuracy: 0.9977\n",
            "Epoch 71/100\n",
            "938/938 [==============================] - 5s 5ms/step - loss: 0.0080 - accuracy: 0.9975\n",
            "Epoch 72/100\n",
            "938/938 [==============================] - 6s 7ms/step - loss: 0.0102 - accuracy: 0.9968\n",
            "Epoch 73/100\n",
            "938/938 [==============================] - 5s 5ms/step - loss: 0.0069 - accuracy: 0.9978\n",
            "Epoch 74/100\n",
            "938/938 [==============================] - 6s 6ms/step - loss: 0.0081 - accuracy: 0.9974\n",
            "Epoch 75/100\n",
            "938/938 [==============================] - 4s 5ms/step - loss: 0.0093 - accuracy: 0.9971\n",
            "Epoch 76/100\n",
            "938/938 [==============================] - 4s 5ms/step - loss: 0.0066 - accuracy: 0.9980\n",
            "Epoch 77/100\n",
            "938/938 [==============================] - 6s 6ms/step - loss: 0.0095 - accuracy: 0.9972\n",
            "Epoch 78/100\n",
            "938/938 [==============================] - 4s 5ms/step - loss: 0.0070 - accuracy: 0.9977\n",
            "Epoch 79/100\n",
            "938/938 [==============================] - 5s 5ms/step - loss: 0.0088 - accuracy: 0.9974\n",
            "Epoch 80/100\n",
            "938/938 [==============================] - 5s 6ms/step - loss: 0.0070 - accuracy: 0.9978\n",
            "Epoch 81/100\n",
            "938/938 [==============================] - 5s 5ms/step - loss: 0.0076 - accuracy: 0.9975\n",
            "Epoch 82/100\n",
            "938/938 [==============================] - 6s 6ms/step - loss: 0.0078 - accuracy: 0.9977\n",
            "Epoch 83/100\n",
            "938/938 [==============================] - 5s 5ms/step - loss: 0.0074 - accuracy: 0.9977\n",
            "Epoch 84/100\n",
            "938/938 [==============================] - 6s 6ms/step - loss: 0.0085 - accuracy: 0.9974\n",
            "Epoch 85/100\n",
            "938/938 [==============================] - 5s 5ms/step - loss: 0.0079 - accuracy: 0.9976\n",
            "Epoch 86/100\n",
            "938/938 [==============================] - 5s 5ms/step - loss: 0.0072 - accuracy: 0.9978\n",
            "Epoch 87/100\n",
            "938/938 [==============================] - 6s 6ms/step - loss: 0.0061 - accuracy: 0.9980\n",
            "Epoch 88/100\n",
            "938/938 [==============================] - 5s 5ms/step - loss: 0.0066 - accuracy: 0.9980\n",
            "Epoch 89/100\n",
            "938/938 [==============================] - 6s 6ms/step - loss: 0.0099 - accuracy: 0.9970\n",
            "Epoch 90/100\n",
            "938/938 [==============================] - 5s 5ms/step - loss: 0.0066 - accuracy: 0.9979\n",
            "Epoch 91/100\n",
            "938/938 [==============================] - 5s 5ms/step - loss: 0.0038 - accuracy: 0.9989\n",
            "Epoch 92/100\n",
            "938/938 [==============================] - 6s 6ms/step - loss: 0.0093 - accuracy: 0.9973\n",
            "Epoch 93/100\n",
            "938/938 [==============================] - 5s 5ms/step - loss: 0.0062 - accuracy: 0.9981\n",
            "Epoch 94/100\n",
            "938/938 [==============================] - 6s 6ms/step - loss: 0.0093 - accuracy: 0.9974\n",
            "Epoch 95/100\n",
            "938/938 [==============================] - 5s 5ms/step - loss: 0.0064 - accuracy: 0.9980\n",
            "Epoch 96/100\n",
            "938/938 [==============================] - 7s 7ms/step - loss: 0.0071 - accuracy: 0.9979\n",
            "Epoch 97/100\n",
            "938/938 [==============================] - 5s 5ms/step - loss: 0.0071 - accuracy: 0.9977\n",
            "Epoch 98/100\n",
            "938/938 [==============================] - 5s 6ms/step - loss: 0.0058 - accuracy: 0.9982\n",
            "Epoch 99/100\n",
            "938/938 [==============================] - 6s 6ms/step - loss: 0.0058 - accuracy: 0.9981\n",
            "Epoch 100/100\n",
            "938/938 [==============================] - 5s 5ms/step - loss: 0.0091 - accuracy: 0.9974\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<keras.callbacks.History at 0x7fd5f3b21bb0>"
            ]
          },
          "metadata": {},
          "execution_count": 16
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# After Augmentation\n",
        "\n",
        "test_loss, test_acc = model.evaluate(x_test, y_test)\n",
        "print(augmented_images.shape)\n",
        "print('Test accuracy after augmentation:', test_acc)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "vK7lDK2u-wSt",
        "outputId": "2de95387-6a54-430f-bff0-e88479ab07b1"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "313/313 [==============================] - 1s 2ms/step - loss: 0.1687 - accuracy: 0.9812\n",
            "(120000, 28, 28, 1)\n",
            "Test accuracy after augmentation: 0.9811999797821045\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Print Test accuracy after augmentation\n",
        "print('Test accuracy after augmentation:', test_acc*100, \"%\")\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "BoBWQ0ZF76x0",
        "outputId": "7502eb47-97f7-4d21-b859-f11745d9f260"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Test accuracy after augmentation: 98.11999797821045 %\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# plot digit after training\n",
        "fig = plt.figure(figsize=(15, 2))\n",
        "print(augmented_images.shape)\n",
        "for i, image in enumerate(augmented_images[:10]):\n",
        "  \n",
        "    # Create a new subplot for each image\n",
        "    plt.subplot(1, 10, i+1)\n",
        "    plt.xticks([])\n",
        "    plt.yticks([])\n",
        "    plt.grid(False)\n",
        "    plt.imshow(image, cmap=plt.cm.binary)\n",
        "    plt.xlabel(y_train[i])\n",
        "\n",
        "\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 148
        },
        "id": "mlAu_-0J5Ygg",
        "outputId": "289e05a2-3171-4be2-98ee-ca5e00a32277"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "(120000, 28, 28, 1)\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1500x200 with 10 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "swxfIBDu82wF"
      },
      "execution_count": null,
      "outputs": []
    }
  ]
}